The 5-HT(2A/2C) receptor antagonist, ritanserin, was reported to retard the acquisition of conditioned responses (CRs) during classical conditioning of the rabbit's nictitating membrane (NM) response. The present study compared the effects of ritanserin on acquisition of CRs to a tone conditioned stimulus (CS) with that of the 5-HT(2A/2C) receptor antagonist, LY-53,857 and the 5-HT2A selective antagonist, MDL-11,939. All three drugs were injected at equimolar doses of 0.067, 0.67 and 6.7 micromol/kg, SC, 1 h before behavioral testing. Ritanserin and MDL-11,939 retarded CR acquisition to a tone CS, while LY-53,857 had no effect. Control experiments demonstrated that ritanserin (1 micromol/kg), MDL-11,939 (1 micromol/kg) and LY-53,857 (2 micromol/kg) had no effect on baseline responding or non-associative responding to the CS. However, both ritanserin and MDL-11,939 impaired the performance of the unconditioned NM reflex, as measured by a decrease in UR amplitudes on US alone trials, while LY-53,857 had no effect. In previously trained animals, ritanserin robustly impaired the performance of CRs, as measured by a reduced ability of the CS to elicit CRs, while the effects of LY-53,857 and MDL-11,939 were marginal. The retardation of associative learning produced by ritanserin and MDL-11,939 may have been due, at least in part, to their impairment of the NM reflex arc. Since MDL-11,939 is a highly selective 5-HT2A antagonist, the retardation of learning and impairment of UR amplitudes produced by MDL-11,939 and ritanserin may have been due to blockade of the 5-HT2A receptor. The ability of ritanserin and MDL-11,939 to produce effects on learning and performance that were opposite to that of 5-HT(2A/2C) agonists suggests that they may be acting as inverse agonists at that receptor. These results stress the importance of the serotonergic system for optimal associative learning and motor function.